Fisheyes After Sanding: Is a Shellac Barrier Worth Testing?

The sanded tabletop looked uniformly dull. Then the fresh finish drew back into little craters. Repeating the same sanding pass may remove the crater rims, but it will not necessarily remove the condition that made the liquid pull away.

On ordinary wooden furniture being deliberately refinished, a dewaxed-shellac barrier can be worth testing when the chosen finish system permits it. It is not a universal fisheye cure. The barrier must bond to what is underneath, remain continuous, and accept the intended topcoat. A smooth first coat answers only part of that question.

First establish what failed

Photograph the wet defect without brushing back through it repeatedly. Note whether it appeared immediately on application or developed later, and whether it follows the same places as the previous failed coat. Keep the product names, mixing details and preparation history with those observations.

Contamination is a possibility, not a diagnosis from a circular mark alone. General Finishes discusses wax and silicone as likely sources of larger craters in its water-based spraying troubleshooting guide, while also describing application conditions that affect leveling. Bubbles, dust nibs and a cloudy film call for different investigations. A nib stands above the film; a crater is a depression where the coating has not formed the intended surface.

Look beyond the tabletop. Used polishing cloths, furniture sprays, applicators and the finishing area may reintroduce the problem. Do not pour suspect working finish back into a clean original container. Set aside the affected materials while you establish a clean trial setup.

Is there a sound surface to isolate?

Let the failed coating reach the removal or sanding condition specified for that product. A soft film that smears under the abrasive is not ready for a controlled leveling step. Remove the failed material by a method compatible with the underlying wood and finish; do not keep grinding each crater until it disappears into the veneer.

An intact but contaminated old finish presents a different problem from one that is peeling or lifting. A barrier laid over a weak coating still depends on that weak coating. Loose finish needs an appropriate removal plan, not another layer to conceal it. Valuable furniture, fragile veneer and uncertain historic finishes deserve specialist assessment before solvents or abrasion change them.

General Finishes' silicone-contamination guidance discusses cleaning and a dewaxed-shellac barrier, but warns that contamination may persist even after stripping. Treat isolation as a trial repair, not a promise that the wood has been decontaminated.

Test the whole finish stack

Verify the exact dewaxed product against both the remaining surface and the proposed topcoat. Ordinary wax-containing shellac is not automatically an equivalent substitute. Check the topcoat's current directions as well as the sealer's: a broad claim on one container does not settle the requirements of the other.

Choose a small, representative area that can be repaired if the test fails. The pristine underside of a table may not have the same polish history as its top, so success there is weak evidence about the contaminated face. A matching offcut can test the new products together, but cannot reproduce an unknown old contamination problem.

Use the verified cleaning, drying and barrier application schedule in that area, then apply the intended topcoat. Judge two separate outcomes:

• Does the liquid form an even film without pulling into craters?
• After the prescribed cure, does the full coating stack pass the finish supplier's adhesion check?

Lifting, softening, wrinkles or renewed craters are reasons to stop the trial. Do not repair a failed test by adding successive unidentified sealers. Ask the finish supplier whether removal and a different system remain viable; persistent contamination may make the desired finish impractical.

Preserve the barrier when smoothing

A barrier is not a sacrificial layer to sand aggressively flat. Any required abrasion must leave the intended isolation coverage intact, especially at edges and profiles. If the schedule calls for 400-grit paper to remove minor roughness, 400-grit sheets in a 10-pack are a cut-to-size option for that limited task. They are not a silicone-removal treatment, and a 400-grit sheet is not automatically interchangeable with a sanding pad bearing another numbering system.

Clean away residue with the method compatible with the next coat. Do not assume that a freshly sanded appearance proves either cleanliness or adhesion. Use the coating safety data for ventilation, ignition control and personal protection; improvised solvent mixtures add another unknown to an already difficult repair.

This is a wood-refinishing decision. Do not transfer a shellac isolation experiment to an automotive basecoat/clearcoat repair or another coating system that has not approved it. The useful outcome is a compatible, adherent finish—not simply a coat that briefly looks free of fisheyes.